Regulatory toxicology in the 21st century is faced with the challenge of having to replace its use of experimental animals in chemical risk assessment with alternative methods. This is due to the introduction of the REACH legislation and the seventh amendment to the cosmetics directive. Such alternative methods include the use of in vitro (cell culture/tissue etc.), in chemico (chemical experiments e.g. determination of reactivity) and in silico (computational) approaches. Importantly, it is envisaged that data from all these alternative sources will be required for the prediction of the animal-based endpoints used in regulatory toxicology. One of the key computational approaches used for data gap filling is category formation and read-across. When using this approach to assess the potential toxicity of a chemical, a chemical category is best defined based on a common molecular initiating event e.g. the formation of a covalent bond with biological nucleophile via the same chemical mechanism. The structural features that define a chemical's membership of such a category can be encoded computationally as structural alerts, which in turn, can be grouped together to form an in silico profiler. The work discussed in this thesis addresses the key shortcoming of traditional in silico profilers, this being that current in silico profilers provided no information about the rate of covalent bond formation for chemicals containing the same structural alert but with different substituents. The research within this thesis addresses this problem through the introduction of a fragment-based approach to in silico profiler development. This fragment-based approach introduces the use of calculated activation energies determined through the use of quantum mechanics calculations which enable chemical reactivity to be predicted. Chapter 3 outlines the development of the approach for α,β-unsaturated aldehydes, ketones and esters which form covalent bonds through Michael addition. Chapter 4 extends the work outlined in Chapter 3 demonstrating how the fragment-based profiler can be used to predict both chemical reactivity and skin sensitisation and toxicity to Tetrahymena pyriformis. Finally, Chapter 5 extends the approach to chemicals capable of reacting with proteins via an SN2 mechanism demonstrating the approach can be applied to any mechanistic domain for which data exist. Overall, this thesis outlines an approach for the development of novel fragment-based in silico profilers capable of quantitatively predicting chemical reactivity and by extension toxicity. It is envisaged that the work outlined in this thesis will be of use primarily in regulatory toxicology, within such tools as the OECD QSAR toolbox.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:762873 |
| Date | January 2018 |
| Creators | Ebbrell, D. J. |
| Contributors | Enoch, S. ; Madden, J. ; Cronin, M. |
| Publisher | Liverpool John Moores University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://researchonline.ljmu.ac.uk/9686/ |
Page generated in 0.0013 seconds